A real-time network-based approach for analysing best–worst data types

Best–worst scaling is a widespread approach in market research used for collecting data on the needs and preferences of people. However, the current preparation of its design and the analysis of the data depends on complex statistical methods. One of the most commonly used models for estimating individual preference probabilities is the hierarchical Bayes model, which can only be applied after the data collection phase. This type of calculation needs more infrastructural background and a large sample to provide accurate estimations. Here, we introduce a new application that enables fast calculations and individual-level real-time estimations, which also has a great potential to ask additional questions depending on the respondent’s answers during live interviews. Our network-based approach (integrating the PageRank algorithm) works well for online surveys, and it supports our dynamic and adaptive, real-time evaluation (DART) of best–worst data types, and results in more relevant decision making in marketing.

Forecasting COVID-19 cases using Machine Learning models

As of April 26, 2020, more than 2,994,958 cases of COVID-19 infection have been confirmed globally, raising a challenging public health issue. A predictive model of the disease would help allocate medical resources and determine social distancing measures more efficiently. In this paper, we gathered case data from Jan 22, 2020 to April 14 for 6 countries to compare different models' proficiency in COVID-19 cases prediction. We assessed the performance of 3 machine learning models including hidden Markov chain model (HMM), hierarchical Bayes model, and long-short-term-memory model (LSTM) using the root-mean-square error (RMSE). The LSTM model had the consistently smallest prediction error rates for tracking the dynamics of incidents cases in 4 countries. In contrast, hierarchical Bayes model provided the most realistic prediction with the capability of identifying a plateau point in the incidents growth curve.

How patrons value casino promotional offers: A conjoint study

Casinos rely extensively on free slot play (FSP) offers for incentivizing patron visitation. However, there has been a lack of understanding its influence on driving patron visitation and patrons’ valuation of FSP compared to other casino promotional offers. This study conducted a conjoint analysis on patrons’ valuations of FSP compared to other promotional offerings at a casino resort. Moreover, this study investigated the roles inter-casino competition and visitation frequency have on patrons’ perceived valuation of FSP through a hierarchical Bayes model. The results show that competition plays a significant negative role on patrons’ valuation of FSP, while competition held insignificant influence on patrons’ valuation of food and beverage (F&B) comp offers. Additionally, patrons who visited the casino more frequently valued FSP greater, while less active patrons valued F&B comp offers more. Using the study’s results, casinos can increase their margins through increased efficiencies with their promotional offering mix.